Back in 1987 Labour fought a brilliant election campaign. They’d recently rebranded with their red rose symbol and all the people I hung out with – twenty-somethings in media, public services, the arts, the NHS, – were convinced that despite what the polls said, Labour was on a roll. The Conservatives won their third term with a swing to Labour of 1.5%.
In 1992 even the polls were in Labour’s favour and we all – late twenties, early thirties, canvassing for Labour in Haringey and Hackney – could tell that this was the end for the Conservatives. John Major won the fourth consecutive term for the Tories because the swing to Labour was under 2%.
Fast forward to last night and the six weeks of campaign preceding it. The polls were all pointing to a hung parliament, but on Twitter the feeling was that the polls were calling it wrong, that Labour was almost certain to be the largest party and would be able to put together an anti-Tory coalition. Yes, well, up to a point Lord Copper.
Because there are two things in play here. The first is that (although the figures are hard to be precise about) Twitter only has around 10 million UK users, who fall mainly in the 18-34 age groups. Now this ain’t bad – it’s considerably more than buy a daily newspaper for example – but it’s not as important as we in media and marketing think it is, especially if we also factor in that many of those ten million aren’t that engaged with the channel.
The second is more significant and that is the ‘echo chamber’ effect. We follow people we know, we find amusing or whose views are similar to ours. Occasionally we’ll engage in debate with someone of a differing opinion, or dive into the timeline of a ‘troll’ (i.e. someone we really don’t agree with), or retweet with comments something we find outrageous, but generally speaking we don’t want unpleasantness in our timeline. In the same way that my group of 1987 friends convinced ourselves of an incorrect outcome, so this time the group of like-minded strangers convinced itself that something different was happening to what actually occurred.
Two conclusions come from this, neither of which are original but which are worth restating – Twitter is not public opinion, it is the views of a section of the population who want to share their views, a self-selecting group of a self-selecting group. Second, your Twitter timeline is a subset of this; basing your decisions or predicting outcomes just from what the people you follow are saying is going to lead to nothing but disappointment.